Sample canonical correlation coefficients of high-dimensional random vectors with finite rank correlations
arXiv:2102.03297
Abstract
Consider two random vectors and of the forms and , where , and are independent vectors with i.i.d. entries of mean 0 and variance 1, and are and deterministic covariance matrices, and and are and deterministic matrices. With independent observations of , we study the sample canonical correlations between and . We consider the high-dimensional setting with finite rank correlations. Let be the squares of the nontrivial population canonical correlation coefficients, and let be the squares of the sample canonical correlation coefficients. If the entries of , and are i.i.d. Gaussian, then the following dichotomy has been shown in [7] for a fixed threshold : for , if , then converges to the right-edge of the limiting eigenvalue spectrum of the sample canonical correlation matrix; if , then converges to a deterministic limit determined by . In this paper, we prove that these results hold universally under the sharp fourth moment conditions on the entries of and . Moreover, we prove the results in full generality, in the sense that they also hold for near-degenerate 's and for 's that are close to the threshold .
arXiv commentBernoulli Journal (to appear)
References in corpus (9)
- Level-Spacing Distributions and the Airy Kernel
- The largest eigenvalue of rank one deformation of large Wigner matrices
- The largest eigenvalues of finite rank deformation of large Wigner matrices: convergence and nonuniversality of the fluctuations
- Universality of covariance matrices
- Averaging Fluctuations in Resolvents of Random Band Matrices
- The outliers of a deformed Wigner matrix
- The largest eigenvalues of sample covariance matrices for a spiked population: diagonal case
- Sample canonical correlation coefficients of high-dimensional random vectors: local law and Tracy-Widom limit
- How to reduce dimension with PCA and random projections?